动态环境下多媒体视频兴趣区的自动捕获系统设计

摘要

When the multimedia video region of interest(ROI)is captured in dynamic environment,the available target capture method can’t acquire the clear image in a specific background environment due to the continuously changing image background and rough layout,so a capture algorithm based on optimal target is proposed to design the automatic capture system for multimedia video ROI in dynamic environment. In the system,the multimedia video image data is acquired by means of the multimedia video image acquisition module,transformed with data format conversion module,and then the converted multime⁃dia video image data can be stored in data cache module. The Gaussian mixture model is adopted in the image algorithm analy⁃sis module to improve the transparency of image background,and capture the clear video ROI image. The captured target image is sent to the core chip ADV7123 in the video image display module for display. The experimental results indicate that the de⁃signed system can capture the moving video ROI target accurately,and obtain high resolution target image.%动态环境下对多媒体视频兴趣区进行捕获时，图像背景是持续转变的，存在高低不平的布局问题，导致当前面向特定背景环境的目标捕获方法，无法获取清晰的图像，提出依据优化目标捕获算法，设计动态环境下多媒体视频兴趣区自动捕获系统。该系统通过多媒体视频图像采集模块采集多媒体视频图像的数据，利用数据格式转换模块变换获取多媒体视频图像数据，完成转换的视频图像数据可存储在数据缓存模块中，图像算法分析模块采用高斯混合模型提高图像背景的透明度，捕获清晰的视频兴趣区图像，并将捕获的图像目标传动核心芯片为ADV7123的视频图像显示模块进行显示。实验结果表明，所设计系统可准确捕获运动的视频兴趣区目标，并且目标清晰度较高。